基于SE模块和ResNet的番茄病虫害识别方法

胡文艺; 王洪坤; 杜育佳

doi:10.19998/j.cnki.2095-1795.2022.09.007

基于SE模块和ResNet的番茄病虫害识别方法

doi: 10.19998/j.cnki.2095-1795.2022.09.007

成都理工大学计算机与网络安全学院，四川成都 610059

详细信息

作者简介:
胡文艺，博士，副教授，主要从事模式识别研究E-mail：64151325@qq.com

中图分类号: S126
计量
- 文章访问数: 58
- HTML全文浏览量: 6
- PDF下载量: 14
- 被引次数: 0
出版历程
- 收稿日期: 2022-04-02
- 修回日期: 2022-06-10
- 出版日期: 2022-09-20

Identification Method of Tomato Diseases and Pests Based on SE Module and ResNet

College of Computer and Cyber Security，Chengdu University of Technology，Chengdu Sichuan 610059，China

摘要

摘要:
番茄病虫害是引起番茄减产的重要因素。精确识别病虫害种类是当前国际热点问题之一，有助于及时有效采取针对性的病虫防治办法，减少和避免因番茄减产导致的经济损失。针对传统虫害识别方法存在效率和精确率低的问题，利用Kaggle网站上的Tomato数据集，构建基于压缩和激励（SE）模块的深度残差网络模型（ResNet），优化番茄病虫害识别方法。结果表明：通过Pytorch框架下的迁移学习，改进后的网络模型对番茄病虫害图像的平均识别准确率最高为97.96%；基于SE模块的ResNet网络模型有助于增强特征区分能力，增加模型的通用性和鲁棒性。研究结果对番茄病虫害的及时监测和处理、提高番茄产量具有重要意义。
- 番茄 /
- 病虫害识别 /
- 迁移学习 /
- 压缩和激励模块 /
- 深度残差网络模型 /
- Pytorch
Abstract:
Tomato diseases and pests are leading factors to cause decline of tomato production． Identifying types of diseases and pests accurately is one of current international hot issues, which helps to take timely and effective measures to control diseases and pests, and reduce and avoid economic losses caused by decline of tomato production． Aiming at problems of low efficiency and accuracy of traditional diseaese and pests identification methods, tomato data set on Kaggle website were used to build a deep residual network model（ResNet）based on Squeeze-and-Excitation（SE）module to optimize tomato pests diseases and identification method． Results showed that: through transfer learning under Pytorch framework, average recognition accuracy of improved network model for tomato pests and diseases images was as high as 97.96%; ResNet network model based on SE module helped to enhance ability of distinguishing features, which increased universality and robustness of the model． Results of this study were of great significance for timely monitoring and treatment of tomato diseases and pests and increase of tomato production．
- tomato /
- identification of diseases and pests /
- transfer learning /
- SE module /
- ResNet /
- Pytorch

HTML全文

图 1 卷积神经网络结构

Figure 1. Structure of convolutional neural networks

下载: 全尺寸图片幻灯片

图 2 残差结构模块

Figure 2. Residual structure module

下载: 全尺寸图片幻灯片

图 3 SE结构

Figure 3. SE structure

下载: 全尺寸图片幻灯片

图 4 SE-ResNet网络结构

Figure 4. SE-ResNet structure

下载: 全尺寸图片幻灯片

图 5 番茄病虫害示例

Figure 5. Example of tomato pests and diseases

下载: 全尺寸图片幻灯片

图 6 不同模型的准确率

Figure 6. Accuracy of different models

下载: 全尺寸图片幻灯片

图 7 不同模型的损失值

Figure 7. Loss values for different models

下载: 全尺寸图片幻灯片

表 1 数据统计

Table 1. Data statistics 单位：张

类别	训练集	测试集
健康	1 702	425
细菌斑	1 920	480
早疫病	1 926	481
晚疫病	1 851	463
叶霉病	1 882	470
白粉病	1 827	457
斑枯病	1 745	436
蜘蛛螨	1 741	435
番茄花叶病毒	1 790	448
黄叶卷曲病毒	1 961	490

下载: 导出CSV

表 2 平均准确率对比

Table 2. Comparison of average accuracy

模型	平均准确率/%
ResNet34	94.23
ResNet50	92.07
SE-ResNet34	97.96
SE-ResNet50	97.31

下载: 导出CSV

表 3 特定类别的识别准确率对比

Table 3. Comparison of recognition accuracy for specific categories 单位：%

类别	ResNet34	ResNet50	SE-ResNet34	SE-ResNet50
健康	97.70	90.76	99.46	99.78
细菌斑	98.88	98.41	98.94	99.82
早疫病	98.49	97.50	99.11	99.48
晚疫病	99.79	99.20	100.00	99.31
叶霉病	97.99	99.43	99.26	97.80
白粉病	98.68	98.73	99.43	99.12
斑枯病	98.08	98.19	99.18	98.47
蜘蛛螨	99.18	98.67	99.49	98.73
番茄花叶病毒	99.94	99.83	99.78	99.79
黄叶卷曲病毒	97.48	96.36	99.24	99.84

下载: 导出CSV

参考文献(39)

[1]	Mitigating impacts of COVID-19 on food trade and markets[EB/OL]．（2020-03-31）．http://www.fao.org/news/story/en/item/1268719/icode/
[2]	张志刚,董涛,李瑞云.基于专利分析的中国番茄产业发展研究[J].农业图书情报学刊,2018,30(11):72-76. ZHANG Zhigang,DONG Tao,LI Ruiyun.Research on the development of tomato industry based on patent analysis[J].Journal of Library and Information Science in Agriculture,2018,30(11):72-76.
[3]	赵美佳,邹通,汤泽君,等.番茄营养成分以及国内外加工现状[J].食品研究与开发,2016,37(10):215-218. doi: 10.3969/j.issn.1005-6521.2016.10.053 ZHAO Meijia,ZOU Tong,TANG Zejun,et al.Nutritional ingredient of tomato and processing status in domestic and overseas[J].Food Research and Development,2016,37(10):215-218. doi: 10.3969/j.issn.1005-6521.2016.10.053
[4]	TEICHERT T, MITTERMAYER M A．Text mining for technology monitoring[C]//Proceedings of IEEE Internati-onal Conferen, Engineering Management（IEMC）, 2002: 596-601．
[5]	HANSSEN I M,LAPIDOT M.Major tomato viruses in the mediterranean basin[J].Advances in Virus Research,2012,84:31-66.
[6]	HIARY H A, AHMAD S B, REYALAT M, et al．Fast and accurate detection and classification of plant diseases[J]．International Journal of Computer Applications, 2011．DOI: 10.5120/2183-2754．
[7]	KEERTHI S S,SHEVADE S K.SMO algorithm for least-squares SVM formulations[J].Neural Computation,2003,15(2):487-507.
[8]	许伟栋,赵忠盖.基于卷积神经网络和支持向量机算法的马铃薯表面缺陷检测[J].江苏农业学报,2018,34(6): 1378- 1385. doi: 10.3969/j.issn.1000-4440.2018.06.025 XU Weidong,ZHAO Zhonggai.Potato surface defects detection based on convolution neural networks and support vector machine algorithm[J].Jiangsu Journal of Agricultural Sciences,2018,34(6): 1378- 1385. doi: 10.3969/j.issn.1000-4440.2018.06.025
[9]	徐岩,米强,刘斌,等.基于字典学习的煤与矸石图像特征识别方法[J].山东科技大学学报(自然科学版),2018,37(3):66-72. XU Yan,MI Qiang,LIU Bin,et al.Feature recognition method of coal and gangue image based on dictionary learning[J].Journal of Shandong University of Science and Technology (Natural Science Edition),2018,37(3):66-72.
[10]	李晓振,徐岩,吴作宏,等.基于注意力神经网络的番茄叶部病害识别系统[J].江苏农业学报,2020,36(3):561-568. doi: 10.3969/j.issn.1000-4440.2020.03.005 LI Xiaozhen,XU Yan,WU Zuohong,et al.Recognition system of tomato leaf disease based on attentional neural network[J].Jiangsu Journal of Agricultural Sciences,2020,36(3):561-568. doi: 10.3969/j.issn.1000-4440.2020.03.005
[11]	邹永杰,张永军,秦永彬,等.应用于番茄病虫害检测的HOG特征与LBP特征的结合[J].南京师范大学学报(工程技术版),2019,19(3):21-28. doi: 10.3969/j.issn.1672-1292.2019.03.004 ZOU Yongjie,ZHANG Yongjun,QIN Yongbin,et al.The combination of HOG features with LBP features applied to tomato disease and pest detection[J].Journal of Nanjing Normal University(Engineering and Technology Edition),2019,19(3):21-28. doi: 10.3969/j.issn.1672-1292.2019.03.004
[12]	吕权．基于神经网络的番茄叶部病害识别[D]．保定: 河北农业大学, 2018． LV Quan．Tomato leaf disease recognition based on neural network[D]．Baoding: Hebei Agricultural Universi-ty, 2018．
[13]	JIANG D, LI F, YANG Y, et al．A tomato leaf diseases classification method based on deep learning[C]// 2020 Chinese Control and Decision Conference （CCDC）．2020．DOI: 10.1109/CCDC49329.2020.9164457．
[14]	刘志勇,张丽秀,钟婷婷,等.基于改进leNet-5的番茄病虫害识别的研究[J].赣南师范大学学报,2020,41(6):70-74. LIU Zhiyong,ZHANG Lixiu,ZHONG Tingting,et al.Study on tomato pest identification based on improved leNet-5[J].Journal of Gannan Normal University,2020,41(6):70-74.
[15]	柴帅,李壮举.基于迁移学习的番茄病虫害检测[J].计算机工程与设计,2019,40(6): 1701- 1705. CHAI Shuai,LI Zhuangju.Detection of tomato pests and diseases based on transfer learning[J].Computer Engineering and Design,2019,40(6): 1701- 1705.
[16]	CHENG X,ZHANG Y,CHEN Y,et al.Pest identification via deep residual learning in complex background[J].Computers and Electronics in Agriculture,2017,141:351-356. doi: 10.1016/j.compag.2017.08.005
[17]	GOODFELLOW I, BENGIO Y, COURVILLE A．De-ep Learning[EB/OL]．（2016）．http://www.deeplearning-book.org．
[18]	HUBEL D H,WIESEL T N.Receptive fields and functional architecture of monkey striate cortex[J].The Journal of Physiology,1968(1):215-243.
[19]	KUNIHIKO,FUKUSHIMA.Neocognitron:a self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position[J].Biological Cybernetics,1980,36(4):193-202. doi: 10.1007/BF00344251
[20]	LECUN Y,BOTTOU L,BENGIO Y,et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE,1998(11): 2278- 2324.
[21]	KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[J].Communications of the ACM,2017,60(6):84-90. doi: 10.1145/3065386
[22]	RUSSAKOVSKY O,DENG J,SU H,et al.ImageNet large scale visual recognition challenge[J].International Journal of Computer Vision,2015,115(3):211-252. doi: 10.1007/s11263-015-0816-y
[23]	赵志焱,杨华,胡志伟,等.基于TACNN的玉露香梨叶虫害识别[J].计算机工程与应用,2021,57(9):176-181. doi: 10.3778/j.issn.1002-8331.2003-0014 ZHAO Zhiyan,YANG Hua,HU Zhiwei,et al.Identification model of pests on yuluxiang pear leaves based on TACNN[J].Computer Engineering and Applications,2021,57(9):176-181. doi: 10.3778/j.issn.1002-8331.2003-0014
[24]	李哲,潘维,张赟.基于深度卷积神经网络的植物病虫害识别[J].电子技术与软件工程,2020(23):194-195.
[25]	HE K, ZHANG X, REN S, et al．Deep residual learning for image recognition[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）．
[26]	SZEGEDY C,LIU W,JIA Y,et al.Going deeper with convolutions[J].IEICE Transactions on Fundamentals of Electronics,Communications and Computer Sciences,2014: abs/1409.4842.
[27]	史冰莹,李佳琦,张磊,等.基于CNN的农作物病虫害图像识别模型[J].计算机系统应用,2020,29(6):89-96. SHI Bingying,LI Jiaqi,ZHANG Lei,et al.Recognition model of crop pests and diseases images based on CNN[J].Computer Systems & Applications,2020,29(6):89-96.
[28]	HE K, JIAN S．Convolutional neural networks at constrained time cost: 10.1109/cvpr.2015.7299173[P]．2014．
[29]	SRIVASTAVA R K, GREFF K, SCHMIDHUBER J．Highway networks[EB/OL]．arXiv: 1505.00387, 2015．
[30]	JIE H, LI S, SAMUEL A, et al．Squeeze-and-excitation networks[J]．IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019．DOI: 10.1109/TPAMI.2019.2913372．
[31]	SANCHEZ J,PERRONNIN F,MENSINK T,et al.Image classification with the fisher vector:theory and practice[J].International Journal of Computer Vision,2013,105(3):222-245.
[32]	SHEN L,SUN G,HUANG Q M,et al.Multi-level discriminative dictionary learning with application to large scale image classification[J].IEEE Transactions on Image Processing,2015,24(10): 3109- 3123.
[33]	YANG J, KAI Y, GONG Y, et al．Linear spatial pyramid matching using sparse coding for image classification[C]// 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition （CVPR 2009）．
[34]	陈雪敏.基于SE-ResNet模型的多角度人脸识别系统设计[J].贵阳学院学报(自然科学版),2020,15(4):10-13. doi: 10.3969/j.issn.1673-6125.2020.04.003 CHEN Xuemin.Design of multi-angle face recognition system based on SE-ResNet model[J].Journal of Guiyang College (Natural Science Edition),2020,15(4):10-13. doi: 10.3969/j.issn.1673-6125.2020.04.003
[35]	QIAN L,CHAO W.Within component and between com-ponent multi-kernel discriminating correlation analysis for colour face recognition[J].IET Computer Vision,2017,11(8):663-674. doi: 10.1049/iet-cvi.2016.0294
[36]	HE K, ZHANG X, REN S, et al．Delving deep into rectifiers: surpassing human-level performance on imagenet classification: IEEE Computer Society, 10.1109/ICCV.2015.123[P]．2015．
[37]	程俊华,曾国辉,鲁敦科,等.基于Dropout的改进卷积神经网络模型平均方法[J].计算机应用,2019,39(6): 1601- 1606. doi: 10.11772/j.issn.1001-9081.2018122501 CHENG Junhua,ZENG Guohui,LU Dunke,et al.Improved convolution neural network model averaging method based on Dropout[J].Journal of Computer Applications,2019,39(6): 1601- 1606. doi: 10.11772/j.issn.1001-9081.2018122501
[38]	SONG Guanghui．Research on image labeling method based on transfer learning and deep convolution features[D]．Hangzhou: Zhejiang University, 2017．
[39]	WANG Y, COTTRELL G W．Bikers are like tobacco shops, formal dressers are like suits: recognizing urban tribes with caffe[C]// 2015 IEEE Winter Conference on Applications of Computer Vision．